Class AbstractLinearSGDTrainer<T extends Output<T>,U,V extends AbstractLinearSGDModel<T>>

All Implemented Interfaces:,<TrainerProvenance>, Trainer<T>, WeightedExamples
Direct Known Subclasses:
LinearSGDTrainer, LinearSGDTrainer, LinearSGDTrainer

public abstract class AbstractLinearSGDTrainer<T extends Output<T>,U,V extends AbstractLinearSGDModel<T>> extends AbstractSGDTrainer<T,U,V,LinearParameters>
A trainer for a linear model which uses SGD.

It's an AbstractSGDTrainer operating on LinearParameters, with the bias folded into the features.


 Bottou L.
 "Large-Scale Machine Learning with Stochastic Gradient Descent"
 Proceedings of COMPSTAT, 2010.
  • Constructor Details

    • AbstractLinearSGDTrainer

      protected AbstractLinearSGDTrainer(StochasticGradientOptimiser optimiser, int epochs, int loggingInterval, int minibatchSize, long seed)
      Constructs an SGD trainer for a linear model.
      optimiser - The gradient optimiser to use.
      epochs - The number of epochs (complete passes through the training data).
      loggingInterval - Log the loss after this many iterations. If -1 don't log anything.
      minibatchSize - The size of any minibatches.
      seed - A seed for the random number generator, used to shuffle the examples before each epoch.
    • AbstractLinearSGDTrainer

      protected AbstractLinearSGDTrainer()
      For olcut.
  • Method Details